kn_iris: A Python repository from abirabedinkhan

       kn_iris is a module for eye iris recognition, can implemented as a biometric system through iris.
       In future versions this helps in diagnosing dieases through iris pattern
       
    ## Installation needs before installing package module 
      ```
         python 
         numpy
         opencv-python
         matplotlib
         opencv-contrib-python
         requests
         scikit-image
         scipy
         imutils==0.5.2
      ```  
      - Create a Models directory & in that directory create a file name irisencodings.pickle on your project folder (Models/irisencodings.pickle).
      - Create a irisdataset directory & in that directory put person's eye iris images under person's name directory.
    
    
    ```shell
    
        Project/
        ├── Models/
        │   ├── irisencodings.pickle/                               # train model
        | 
        ├── irisdataset/ 
        │   ├── person1 name/                                       # person1 directory
        |   │   ├── eye iris images of person1 /                    # images of person eye iris
        │   ├── person2 name/                                       # person2 directory
        |   │   ├── eye iris images of person2 /                    # images of person eye iris
        │   ├── person3 name/                                       # person3 directory
        |   │   ├── eye iris images of person3 /                    # images of person eye iris                   
     
    ```
    ## Installation
    
    ##### - Install Python
    
    [Windows](http://timmyreilly.azurewebsites.net/python-flask-windows-development-environment-setup/), [Mac](http://docs.python-guide.org/en/latest/starting/install/osx/), [Linux](https://docs.aws.amazon.com/cli/latest/userguide/awscli-install-linux-python.html)
    
    
    ##### - Install package module using setup.py::
      ```
        $ cd kn_iris/ 
        $ python3 setup.py
      ```  
    #### use package within your project
        clone repository to your project path
        using gitclone:
        
     
    ## Run Project
    
    Once all the settings of project are configured, you are ready to run your project. To start import kn_iris module.
    
    ```shell
       import kn_iris
    ```
    
    After import, need to train existing images and create encoding module once on start :
    
    ```shell
       kn_iris.iris_model_train(dataset_path, encoding_model_path)
       irisdataset_path        ===>  'irisdataset/'
       encoding_model_path  ===>  'Models/irisencodings.pickle'
    ```
    
    Once model is trained then its ready to test with real-time images:
    
    ```shell
       iris_name = kn_iris.iris_model_test(encoding_model_path,real_time_image_path) 
       encoding_model_path   ===>  Models/irisencodings.pickle
       real_time_image_path       ===>  real-time_image_path
       iris_name                  ===>  it returns predicted person name if image matches with trained image model person image & if not then it returns name as unmatch.
    ```
    
    
    ##Requirements
    
      * Need clearer images from the scanner.
      * Images shouldn't capture on direct sunlight.
      * Person shouldn't use glass or lens on eye scanning.
      * All scanned images need to be on same shapes/size(eg - 320x240).
      * As per image size and quality/noise, need to change parameter of filters according.
      * 90% above eye iris need to be capture on image taken from scanner.
      * Need min 5 clearer images to train a model.
      * After all this done according, set threshold of Hamming Distance to recognize.
    
    
    
    ## Support
    
    still a lot need to be implemented open for your contributions
abirabedinkhan/kn_iris
